## 7.10.2 Using Outflow Boundaries

As noted in Section  7.10.1, the outflow boundary condition is obeyed in fully-developed flows where the diffusion flux for all flow variables in the exit direction are zero. However, you may also define outflow boundaries at physical boundaries where the flow is not fully developed--and you can do so with confidence if the assumption of a zero diffusion flux at the exit is expected to have a small impact on your flow solution. The appropriate placement of an outflow boundary is described by example below.

• Outflow boundaries where normal gradients are negligible: Figure  7.10.1 shows a simple two-dimensional flow problem and several possible outflow boundary location choices. Location C shows the outflow boundary located upstream of the plenum exit but in a region of the duct where the flow is fully-developed. At this location, the outflow boundary condition is exactly obeyed.

• Ill-posed outflow boundaries: Location B in Figure  7.10.1 shows the outflow boundary near the reattachment point of the recirculation in the wake of the backward-facing step. This choice of outflow boundary condition is ill-posed as the gradients normal to the exit plane are quite large at this point and can be expected to have a significant impact on the flow field upstream. Because the outflow boundary condition ignores these axial gradients in the flow, location B is a poor choice for an outflow boundary. The exit location should be moved downstream from the reattachment point.

Figure  7.10.1 shows a second ill-posed outflow boundary at location A. Here, the outflow is located where flow is pulled into the FLUENT domain through the outflow boundary. In situations like this the FLUENT calculation typically does not converge and the results of the calculation have no validity. This is because when flow is pulled into the domain through an outflow, the mass flow rate through the domain is "floating'' or undefined. In addition, when flow enters the domain through an outflow boundary, the scalar properties of the flow are not defined. For example, the temperature of the flow pulled in through the outflow is not defined. ( FLUENT chooses the temperature using the temperature of the fluid adjacent to the outflow, inside the domain.) Thus you should view all calculations that involve flow entering the domain through an outflow boundary with skepticism. For such calculations, pressure outlet boundary conditions (see Section  7.8) are recommended.

 Note that convergence may be affected if there is recirculation through the outflow boundary at any point during the calculation, even if the final solution is not expected to have any flow reentering the domain. This is particularly true of turbulent flow simulations.

Previous: 7.10.1 FLUENT's Treatment at
Up: 7.10 Outflow Boundary Conditions
Next: 7.10.3 Mass Flow Split